From Raw EEG and MEG to Event-Related Potentials and Fields

In this unit, you’ll learn how to go from the raw data recorded at EEG and MEG sensors to event-related responses and apply statistical analyses to compare brain responses to different experimental conditions.

In the first lesson, you’ll load raw EEG and MEG data into MNE-Python and explore it via a graphical user interface. You’ll see how to select channels, annotate data and work with different file formats. The second lesson explores how EEG data is generated from sources in the brain. You’ll use a model of the brain to select sources, simulate activity at those sources and project them to the sensors to obtain realistic EEG data. This will give you some intuitions about how EEG recordings relate to the underlying neural activity and provide you with a tool to generate simulations against which you can test your analysis pipeline.

In the third lesson you’ll learn how to divide continuous EEG recordings into epochs around events (e.g. stimuli) and average them to compute and visualize event-related potentials (ERPs). You’ll also explore how changing the EEG reference affects the appearance of the ERP and how we can identify and remove segments contaminated by noise. The last lesson focuses on statistical analysis of evoked response fields (ERFs) computed by epoching and averaging continuous MEG. You’ll learn how to use a t-test to compare the ERFs between experimental conditions sample-by-sample and channel-by-channel and correct for multiple comparisons. Finally, the lesson will introduce the non-parametric permutation cluster test that can test for differences across channels and time-points while avoiding the multiple comparisons problem